population structure of uk biobank and ancient eurasians reveals ...€¦ · 27/05/2016 · we...

1

Population structure of UK Biobank and ancient Eurasians reveals adaptation at genes influencing blood pressure Kevin J. Galinsky1,2, Po-Ru Loh2,3, Swapan Mallick2,4, Nick J. Patterson2, Alkes L. Price1,2,3

1. Department of Biostatistics, Harvard T.H. Chan School of Public Health, Boston, MA

02115, USA

2. Program in Medical and Population Genetics, Broad Institute of MIT and Harvard,

Cambridge, MA 02142, USA

3. Department of Epidemiology, Harvard T.H. Chan School of Public Health, Boston, MA

02115, USA

4. Department of Genetics, Harvard Medical School, Boston, MA 02115, USA

Correspondence should be addressed to K.J.G. ([email protected]) or A.L.P.

([email protected]).

Abstract

Analyzing genetic differences between closely related populations can be a powerful way to

detect recent adaptation. The very large sample size of the UK Biobank is ideal for detecting

selection using population differentiation, and enables an analysis of UK population structure at

fine resolution. In analyses of 113,851 UK Biobank samples, population structure in the UK is

dominated by 5 principal components (PCs) spanning 6 clusters: Northern Ireland, Scotland,

northern England, southern England, and two Welsh clusters. Analyses with ancient Eurasians

show that populations in the northern UK have higher levels of Steppe ancestry, and that UK

certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (which was notthis version posted May 27, 2016. ; https://doi.org/10.1101/055855doi: bioRxiv preprint

https://doi.org/10.1101/055855

2

population structure cannot be explained as a simple mixture of Celts and Saxons. A scan for

unusual population differentiation along top PCs identified a genome-wide significant signal of

selection at the coding variant rs601338 in FUT2 (𝑝 = 9.16 × 10−9). In addition, by combining

evidence of unusual differentiation within the UK with evidence from ancient Eurasians, we

identified new genome-wide significant (𝑝 < 5 × 10−8) signals of recent selection at two

additional loci: CYP1A2/CSK and F12. We detected strong associations to diastolic blood

pressure in the UK Biobank for the variants with new selection signals at CYP1A2/CSK (𝑝 =

1.10 × 10−19) and for variants with ancient Eurasian selection signals in the ATXN2/SH2B3

locus (𝑝 = 8.00 × 10−33), implicating recent adaptation related to blood pressure.

Introduction

Detecting signals of selection can provide biological insights into adaptations that have shaped

human history1–4. Searching for genetic variants that are unusually differentiated between

populations is a powerful way to detect recent selection5; this approach has been applied to

detect signals of selection linked to lactase resistance6,7, fatty acid decomposition8, hypoxia

response9–11, malaria resistance12–14, and other traits and diseases15–18.

Leveraging population differentiation to detect selection is particularly powerful when

analyzing closely related subpopulations with large sample sizes19. Here, we analyze 113,851

samples of UK ancestry from the UK Biobank (see URLs) in conjunction with recently published

People of the British Isles (PoBI)20 and ancient DNA21–24 data sets to draw inferences about

population structure and recent selection. We employ a recently developed selection statistic


https://doi.org/10.1101/055855

3

that detects unusual population differentiation along continuous principal components (PCs)

instead of between discrete subpopulations25, and combine our results with independent

results from ancient Eurasians23. We detect three new signals of selection, and show that

genetic variants with both new and previously reported23 signals of selection are strongly

associated to diastolic blood pressure in UK Biobank samples.

Results

Population Structure in the UK Biobank

We restricted our analyses of population structure to 113,851 UK Biobank samples of UK

ancestry and 202,486 SNPs after quality control (QC) filtering and linkage disequilibrium (LD)

pruning (see Online Methods). We ran principal components analysis (PCA) on this data, using

our FastPCA implementation25 (see URLs). We determined that the top 5 PCs represent

geographic population structure (Figure 1), by visually examining plots of the top 10 PCs

(Supplementary Figure 1), observing that the eigenvalues for the top 5 PCs were above

background levels, and that the eigenvectors were correlated with birth coordinate

(Supplementary Table 1). The eigenvalue for PC1 was 20.99, which corresponds to the

eigenvalue that would be expected at this sample size for two discrete subpopulations of equal

size with an 𝐹𝑆𝑇 of 1.76 × 10−4 (Supplementary Table 1).

We ran k-means clustering on these 5 PCs to partition the samples into 6 clusters, since 𝐾 PCs

can differentiate 𝐾 + 1 populations (Figure 1, Table 1, Supplementary Figure 2). To identify the

populations underlying the 6 clusters, we projected the PoBI dataset20, comprising 2,039


https://doi.org/10.1101/055855

4

samples from 30 regions of the UK, onto the UK Biobank PCs (Figure 2, Supplementary Figure

3). The individuals in the PoBI study were from rural areas of the UK and had all four

grandparents born within 80 km of each other, allowing a glimpse into the genetics of the UK

before the increase in mobility of the 20th century. We selected representative PoBI sample

regions that best aligned with the 6 UK Biobank clusters by comparing centroids of each

projected population region with those from the UK Biobank clusters via visual inspection (see

Online Methods, Table 1). The largest cluster represented southern England, three clusters

represented different regions in the northern UK (northern England, Northern Ireland and

Scotland) and two clusters represented north and south Wales. The PCs separated the six UK

clusters along two general geographical axes: a north-south axis and a Welsh-specific axis. PC1

and PC3 both separated individuals on north-south axes of variation, with southern England on

one end and one of the northern UK clusters on the other. PC2 separated the Welsh clusters

from the rest of the UK. PC4 separated the Scotland cluster from the Northern Ireland cluster.

PC5 separated the north Wales and south Wales (also known as Pembrokeshire) clusters from

each other.

We next analyzed UK Biobank population structure in conjunction with ancient DNA samples.

Modern European populations are known to have descended from three ancestral populations:

Steppe, Mesolithic Europeans and Neolithic farmers21,22. We projected ancient samples from

these three populations as well as ancient Saxon samples24 onto the UK Biobank PCs (Figure 3,

Supplementary Figure 4, see Online Methods). These populations were primarily differentiated

along PC1 and PC3, indicating higher levels of Steppe ancestry in northern UK populations.


https://doi.org/10.1101/055855

5

Additionally, the lack of any ancient sample correlation with PC2 suggests that Welsh

populations are not differentially admixed with any ancient population in our data set, and

likely underwent Welsh-specific genetic drift. We confirmed these findings by projecting pan-

European POPRES26 samples onto the UK Biobank PCs (see Online Methods, Supplementary

Figure 5) noting that of the continental European populations, Russians (who have the most

Steppe ancestry) lie on one side and Spanish and Italians (who have least)22 lie on the other side

along PC1 and PC3, and that none of the continental European populations projected onto the

same regions as the Welsh on PC2 and PC5.

In addition to the impact of ancient Eurasian populations, we know that the genetics of the UK

has been strongly impacted by Anglo-Saxon migrations since the Iron Age24, with the Angles

arriving in eastern England and the Saxons in southern England. The Anglo-Saxons interbred

with the native Celts, which explains much of the genetic landscape in the UK. We analyzed a

variety of samples from Celtic (Scotland and Wales) and Anglo-Saxon (southern and eastern

England) populations from modern Britain in conjunction with the PoBI samples20 and 10

ancient Saxon samples from eastern England24 in order to assess the relative amounts of Steppe

ancestry. We computed 𝑓4 statistics27 of the form 𝑓4(𝑆𝑡𝑒𝑝𝑝𝑒, 𝑁𝑒𝑜𝑙𝑖𝑡ℎ𝑖𝑐 𝐹𝑎𝑟𝑚𝑒𝑟; 𝑃𝑜𝑝1, 𝑃𝑜𝑝2),

where Steppe and Neolithic Farmer populations are from ref. 21,22, Pop1 is either a modern

Celtic or ancient Saxon population and Pop2 is a modern Anglo-Saxon population (Table 2,

Supplementary Table 2). This statistic is sensitive to Steppe ancestry with positive values

indicating more Steppe ancestry in Pop1 than Pop2. We consistently obtained significantly

positive 𝑓4 statistics, implying that both the modern Celtic samples and the ancient Saxon


https://doi.org/10.1101/055855

6

samples have more Steppe ancestry than the modern Anglo-Saxon samples from southern and

eastern England. This indicates that southern and eastern England is not exclusively a genetic

mix of Celts and Saxons. There are a variety of possible explanations, but one is that the present

genetic structure of Britain, while subtle, is quite old, and that southern England in Roman

times already had less Steppe ancestry than Wales and Scotland.

Signals of Natural Selection

We searched for signals of selection using a recently developed selection statistic that detects

unusual population differentiation along continuous PCs25. Notably, this statistic is able to

detect selection signals at genome-wide significance. We analyzed the top 5 UK Biobank PCs

(which were computed using LD-pruned SNPs), and computed selection statistics at 510,665

SNPs, reflecting the set of SNPs after QC but before LD-pruning (see Online Methods). The

Manhattan plot for PC1 is reported in Figure 4, with additional plots in Supplementary Figure 6.

We detected genome-wide significant signals of selection at FUT2 and at several loci with

widely known signals of selection (Table 3). Loci with suggestive signals of selection (𝑝 < 10−6)

are reported in Supplementary Table 3. FUT2 has also previously been reported as a target of

natural selection28,29, although those results focused on frequency differences between highly

diverged continental populations whereas our results implicate much more recent selection.

FUT2 encodes fucosyltransferase 2, an enzyme that affects the Lewis blood group. The SNP

with the most significant p-value, rs601338, is a coding variant where the variant rs601338*G

encodes the secretor allele and the rs601338*A variant encodes the nonsecretor allele, which

protects against the Norwalk norovirus30,31. This SNP also affects the progression of HIV

infection32, and is associated with vitamin B12 levels33, Crohn’s disease34, celiac disease and


https://doi.org/10.1101/055855

7

inflammatory bowel disease35, possibly due to changes in gut microbiome energy metabolism36.

rs601338*A is more common in northern UK samples (Supplementary Table 4). Similar allele

frequency patterns were also observed in GERA37 and PoBI20 samples at rs492602 and rs676388

(Supplementary Table 4), two linked SNPs in FUT2 whose allele frequencies vary on a north-

south axis in UK Biobank data. rs492602 and rs676388 were suggestively significant (𝑝 <

1.00 × 10−6) but not genome-wide-significant in tests for selection using the GERA data set

(Supplementary Table 5), emphasizing the advantage of analyzing more closely related

subpopulations in very large sample sizes in the UK Biobank data set. These three SNPs were

also significant when analyzing the 6 UK Biobank clusters described above using a test for

selection based on unusual differentiation between discrete subpopulations (Supplementary

Table 6).

To detect additional signals of selection, we combined our PC-based selection statistics from

the UK Biobank data with a previously described selection statistic that detects unusual allele

frequency differences after the admixture of ancient Eurasian populations by identifying SNPs

whose allele frequencies are inconsistent with admixture proportions inferred from genome-

wide data23. For each of PC1-PC5 in UK Biobank, we summed our chi-square (1 d.o.f.) selection

statistics for that PC with the chi-square (4 d.o.f.) selection statistics from ref. 23 to produce

chi-square (5 d.o.f.) statistics that combine these independent signals (see Online Methods).

We confirmed the independence of the two selection statistics by checking that the combined

statistics were not inflated, as well as by examining the correlations between the two selection

statistics (Supplementary Table 7). We looked for signals that were genome-wide significant in


https://doi.org/10.1101/055855

8

the combined selection statistic but not in either of the constituent UK Biobank or ancient

Eurasian selection statistics. Results are reported in Table 4.

We detected new genome-wide significant signals of selection at the F12 and CYP1A2/CSK loci.

We are not currently aware of previous evidence of selection at F12. F12 codes for coagulation

factor XII, a protein involved in blood clotting38. The SNP at the F12 locus, rs2545801 was

suggestively significant in the ancient Eurasian analysis (𝑝 = 5.35 × 10−8), and combining it

with the UK Biobank selection statistic on PC2 produced a genome-wide significant signal. This

SNP has been associated with activated partial thromboplastin time, a measure of blood

clotting speed where shorter time is a risk factor for strokes39. An additional significant SNP at

F12, rs2731672, affects expression of F12 in liver40 and is associated with plasma levels of factor

XII41. The CYP1A2/CSK locus has previously been reported as a target of natural selection when

comparing inter-continental allele and haplotype frequencies42,43, but our results implicate

much more recent selection. The two detected SNPs at this locus are in strong LD (𝑟2 = 0.858).

The top SNP, rs1378942, is in an intron in the CSK gene. This SNP has greatly varying allele

frequency across continents43, is associated with blood pressure44,45 and systemic sclerosis (an

autoimmune disease affecting connective tissue)46. The second SNP, rs2472304 in CYP1A2, is

associated with esophageal cancer47, caffeine consumption48 and may mediate the protective

effect of caffeine on Parkinson’s disease49.

We tested SNPs with genome-wide significant signals of selection in the constituent UK Biobank

or ancient Eurasian scans or the combined scan for association with 15 phenotypes in the UK


https://doi.org/10.1101/055855

9

Biobank data set, using the top 5 PCs as covariates (Supplementary Table 8, see Online

Methods). The top SNP at F12 (rs2545801) was associated with height (𝑝 = 4.8 × 10−11), and

the top SNP at CYP1A2/CSK (rs1378942) was associated with diastolic blood pressure (DBP)

(𝑝 = 3.6 × 10−19) and hypertension (𝑝 = 4.8 × 10−9), consistent with previous findings50. We

detected additional associations with DBP (𝑝 = 8.00 × 10−33) and hypertension (𝑝 = 1.30 ×

10−9) at the ATXN2/SH2B3 locus which was reported as under selection in the ancient Eurasian

scan. The top SNP in ATXN2/SH2B3, rs3184504, is known to be associated with blood

pressure51. We note that PC1 and PC3 were strongly associated with height in the UK Biobank

data set, and PC3 and PC4 were associated with DBP(Supplementary Table 9). GRK452, AGT52

and ATP1A114 have also been reported to be under selection and to be associated with DBP or

hypertension. None of the SNPs in GRK4 or ATP1A1 were found to be under selection or

associated with DBP or hypertension in our analyses. The AGT SNP rs699 was associated with

DBP (𝑝 = 7.2 × 10−10) and nominally associated to hypertension (𝑝 = 4.8 × 10−4), although it

did not produce a significant signal of selection in our analyses.

Discussion

In this study, we used PCA to analyze the population structure of a large UK cohort (𝑁 =

113,851). We detected 5 PCs representing geographic population structure that partitioned

this cohort into six subpopulation clusters. Projecting ancient samples onto these PCs revealed

greater Steppe ancestry in northern UK samples. No ancient samples were found to vary along

the Welsh-specific axis, suggesting that the Welsh populations differ from the rest of the UK


https://doi.org/10.1101/055855

10

due to drift and not different levels of admixture. We also determined that UK population

structure cannot be explained as a simple mixture of Celts and Saxons.

We leveraged the subtle population structure and large sample size of the UK Biobank data set

to detect signals of natural selection. We determined that the rs601338*A allele of FUT2 was

more common in northern UK samples, suggesting that pathogens may have exerted selective

pressure in those populations. Combining a selection statistic that detects selection via

population differentiation within the UK with a separate statistic that detects selection since

ancient population admixture in Europe, we were able to detect selection at two additional loci,

F12 and CYP1A2/CSK. We additionally found associations to diastolic blood pressure at

CYP1A2/CSK and at the ATXN2/SH2B3 locus implicated in a previous selection scan.

We conclude by noting three limitations in our work. First, we employed PCA, a widely used

method for analyzing population structure25,53,54, but haplotype-based methods such as

fineSTRUCTURE may be more powerful20,55,56; recent advances in computationally efficient

phasing57,58 increase the prospects for applying such methods to biobank scale data. Second,

we employed methods designed to detect selection at individual loci, but did not employ

methods to detect polygenic selection59–63; our observation that top PCs were correlated with

height and DBP in the UK Biobank data set, which could potentially be consistent with the

action of polygenic selection on these traits, motivates further analyses of possible polygenic

selection. Finally, the PC-based test for selection that we employed assumes that allele

frequencies vary linearly along a PC. The spatial ancestry analysis (SPA) method64–66 allows for a


https://doi.org/10.1101/055855

11

logistic relationship between allele frequency and ancestry, and is not constrained by this

limitation. However, the advantage of the PC-based test for selection is that it allows for the

detection of genome-wide significant signals, a key consideration in genome scans for selection.

Online Methods

UK Biobank data set

The UK Biobank phase 1 data release contains 847,131 SNPs and 152,729 samples. We

removed SNPs that were multi-allelic, had a genotyping rate less than 99%, or had minor allele

frequency (MAF) less than 1%. We also removed samples with non-British ancestry as well as

samples with a genotyping rate less than 98%. This left 510,665 SNPs and 118,650 samples, a

data set that we call “QC*.” Using PLINK267 (see URLs), we removed SNPs not in Hardy-

Weinberg equilibrium (𝑝 < 10−6), and we LD-pruned SNPs to have 𝑟2 < 0.2. We then

generated a genetic relationship matrix (GRM) and removed one of each any pair of samples

with relatedness greater than 0.05. This data set, which we call “LD,” contained 210,113 SNPs

and 113,851 samples. Taking the full set of SNPs from the QC* data set and the set of

unrelated samples from the LD data set produces the final “QC” dataset.

PoBI and POPRES data sets

The 2,039 UK PoBI samples were a subset of the 4,371 samples collected as part of the PoBI

project20. The 2,039 samples were a subset of the 2,886 samples genotyped on the Illumina

Human 1.2M-Duo genotyping chip, with 2,510 samples passing QC procedures and 2,039

samples with all four grandparents born within 80km of each other. We also examined 2,988


https://doi.org/10.1101/055855

12

European POPRES samples from the LOLIPOP and CoLaus collections26. These samples were

genotyped on the Affymetrix GeneChip 500K Array.

Ancient DNA data sets

Ancient DNA was gathered from several regions. 9 Steppe samples were collected from the

Yamna oblast in Russia22, 7 west-European hunter-gatherers from Loschbour21, 26 Neolithic

farmer samples from the Anatolian region22, and 10 Saxon samples from three sites in the UK24.

DNA was extracted from bone tissue, PCR amplified and then purified using a hybrid capture

approach22–24. The resulting DNA was sequenced on Illumina MiSeq, HiSeq or NextSeq

platforms. Sequenced reads were aligned to the human genome using BWA and called SNPs

were intersected with the SNPs found on the Human Origins Array27.

PCA

We ran PCA on the UK Biobank LD dataset using the FastPCA software in EIGENSOFT25 (see

URLs). We identified several artifactual PCs that were dominated by regions of long-range LD

(Supplementary Figure 7). Removing loci with significant or suggestive selection signals

(Supplementary Table 10) along with their flanking 1Mb regions from the LD data set and

rerunning PCA eliminated these artifactual PCs (Supplementary Figure 1). We refer to the

resulting data set with 202,486 SNPs and 113,851 samples as the “PC” dataset.

PC Projection

We projected PoBI20 (642,288 SNPs, 2,039 samples from 30 populations), POPRES26 (453,442

SNPs, 4,079 samples from 60 populations) and ancient DNA22,23 (159,588 SNPs, 52 samples from

4 populations) samples onto the UK Biobank PCs via PC projection53. The SNPs in the UK

Biobank QC data set were intersected with those in the projected data set and A/T and C/G


https://doi.org/10.1101/055855

13

SNPs were removed due to strand ambiguity (75,254, 37,593 and 24,467 SNPs for PoBI, POPRES

and ancient DNA, respectively). The intersected set of SNPs was stringently LD-pruned for 𝑟2 <

0.05 using PLINK267 (see URLs) (leaving 27,769, 20,914 and 15,722 SNPs respectively). SNP

weights were computed for the intersected set of SNPs and these weights were then used to

project the new samples onto the UK Biobank PCs53.

PCA-based selection statistic

PCA is equivalent to the singular value decomposition (𝑿 = 𝑼𝚺𝑽𝑇) where 𝑿 is the normalized

genomic matrix, 𝑼 is the matrix of left singular vectors, 𝑽 is the matrix of right singular vectors,

and 𝚺 is a diagonal matrix of singular values. The singular values are related to the eigenvalues

of the genetic relationship matrix (GRM) by the relationship 𝚲 = 𝚺2/𝑀, where 𝑀 is the number

of SNPs used to compute the GRM 𝑿𝑇𝑿/𝑀. The matrix 𝑼 has the properties 𝑼𝑇𝑼 = 𝑰 and 𝑼 =

𝑿𝑽𝚺−1. By the central limit theorem, the elements of 𝑼 follow a normal distribution and after

rescaling by 𝑀 they follow a chi-square (1 d.o.f.) distribution. In other words, the statistic

𝑀(𝑿𝑖𝑽𝑘)2/Σ𝑘2 = (𝑿𝑖𝑽𝑘)/𝚲𝑘 for the ith SNP at the kth PC follows a chi-square (1 d.o.f.)

distribution25. One benefit of this statistic is that the PCs can be generated on one set of SNPs

(here we used the PC dataset described earlier) and the selection statistic can be calculated on

another set of SNPs (we used the QC dataset).

Signals of selection were clustered by considering all SNPs for which the 𝑝-value along at least

one PC was less than an initial threshold (which we set at 10−6) and clustering together SNPs

within 1Mb. We defined genome-wide significant loci based on clusters that contained at least

one SNP with a 𝑝-value smaller than the genome-wide significance threshold. Since we

analyzed 5 PCs and 510,665 SNPs, the genome-wide significance threshold was


https://doi.org/10.1101/055855

14

0.05/(5 × 510,665) = 1.96 × 10−8. We defined suggestive loci based on clusters with at

least two SNPs crossing the initial threshold (but none crossing the genome-wide significance

threshold).

Combined selection statistic

We intersected the chi-square (4 d.o.f.) ancient Eurasian selection statistics for 1,004,613 SNPs

from Mathieson et al.23 with the PC-based chi-square (1 d.o.f.) UK Biobank selection statistics

for 510,665 QC SNPs, producing a list of 115,066 SNPs. For each SNP and each PC, we added

the ancient Eurasian selection statistics to the UK Biobank selection statistics for that PC,

producing chi-square (5 d.o.f.) statistics which we corrected using genomic control.

Association tests

Association analyses were performed using PLINK267 with the top 5 PC as covariates using the “-

-linear” or “--logistic” flags.

Acknowledgments

We thank Iain Mathieson and David Reich for helpful discussions and Stephan Schiffels for

technical assistance with Saxon samples. This research was conducted using the UK Biobank

Resource and was funded by NIH grant R01 HG006399.

URLs

UK Biobank: http://www.ukbiobank.ac.uk/

EIGENSOFT v6.1.1 (FastPCA and PC-based selection statistic):

http://www.hsph.harvard.edu/alkes-price/software/


http://www.ukbiobank.ac.uk/

http://www.hsph.harvard.edu/alkes-price/software/

https://doi.org/10.1101/055855

15

PLINK2: https://www.cog-genomics.org/plink2

References

1. Sabeti, P. C. et al. Detecting recent positive selection in the human genome from haplotype

structure. Nature 419, 832–837 (2002).

2. Nielsen, R., Hellmann, I., Hubisz, M., Bustamante, C. & Clark, A. G. Recent and ongoing

selection in the human genome. Nat. Rev. Genet. 8, 857–868 (2007).

3. Novembre, J. & Di Rienzo, A. Spatial patterns of variation due to natural selection in humans.

Nat. Rev. Genet. 10, 745–755 (2009).

4. Scheinfeldt, L. B. & Tishkoff, S. A. Recent human adaptation: genomic approaches,

interpretation and insights. Nat. Rev. Genet. 14, 692–702 (2013).

5. Shriver, M. D. et al. The genomic distribution of population substructure in four populations

using 8,525 autosomal SNPs. Hum. Genomics 1, 274 (2004).

6. Bersaglieri, T. et al. Genetic Signatures of Strong Recent Positive Selection at the Lactase

Gene. Am. J. Hum. Genet. 74, 1111–1120 (2004).

7. Tishkoff, S. A. et al. Convergent adaptation of human lactase persistence in Africa and

Europe. Nat. Genet. 39, 31–40 (2007).

8. Fumagalli, M. et al. Greenlandic Inuit show genetic signatures of diet and climate adaptation.

Science 349, 1343–1347 (2015).

9. Yi, X. et al. Sequencing of 50 Human Exomes Reveals Adaptation to High Altitude. Science

329, 75–78 (2010).


https://www.cog-genomics.org/plink2

https://doi.org/10.1101/055855

16

10. Bigham, A. et al. Identifying Signatures of Natural Selection in Tibetan and Andean

Populations Using Dense Genome Scan Data. PLoS Genet 6, e1001116 (2010).

11. Lorenzo, F. R. et al. A genetic mechanism for Tibetan high-altitude adaptation. Nat.

Genet. 46, 951–956 (2014).

12. Hamblin, M. T. & Di Rienzo, A. Detection of the Signature of Natural Selection in

Humans: Evidence from the Duffy Blood Group Locus. Am. J. Hum. Genet. 66, 1669–1679

(2000).

13. Ayodo, G. et al. Combining Evidence of Natural Selection with Association Analysis

Increases Power to Detect Malaria-Resistance Variants. Am. J. Hum. Genet. 81, 234–242

(2007).

14. Gurdasani, D. et al. The African Genome Variation Project shapes medical genetics in

Africa. Nature 517, 327–332 (2015).

15. Lamason, R. L. et al. SLC24A5, a Putative Cation Exchanger, Affects Pigmentation in

Zebrafish and Humans. Science 310, 1782–1786 (2005).

16. Perry, G. H. et al. Diet and the evolution of human amylase gene copy number variation.

Nat. Genet. 39, 1256–1260 (2007).

17. Hancock, A. M. et al. Adaptations to Climate-Mediated Selective Pressures in Humans.

PLoS Genet 7, e1001375 (2011).

18. Ko, W.-Y. et al. Identifying Darwinian Selection Acting on Different Human APOL1

Variants among Diverse African Populations. Am. J. Hum. Genet. 93, 54–66 (2013).


https://doi.org/10.1101/055855

17

19. Bhatia, G. et al. Genome-wide Comparison of African-Ancestry Populations from CARe

and Other Cohorts Reveals Signals of Natural Selection. Am. J. Hum. Genet. 89, 368–381

(2011).

20. Leslie, S. et al. The fine-scale genetic structure of the British population. Nature 519,

309–314 (2015).

21. Lazaridis, I. et al. Ancient human genomes suggest three ancestral populations for

present-day Europeans. Nature 513, 409–413 (2014).

22. Haak, W. et al. Massive migration from the steppe was a source for Indo-European

languages in Europe. Nature 522, 207–211 (2015).

23. Mathieson, I. et al. Genome-wide patterns of selection in 230 ancient Eurasians. Nature

528, 499–503 (2015).

24. Schiffels, S. et al. Iron Age and Anglo-Saxon genomes from East England reveal British

migration history. Nat. Commun. 7, 10408 (2016).

25. Galinsky, K. J. et al. Fast Principal-Component Analysis Reveals Convergent Evolution of

ADH1B in Europe and East Asia. Am. J. Hum. Genet. 98, 456–472 (2016).

26. Nelson, M. R. et al. The Population Reference Sample, POPRES: A Resource for

Population, Disease, and Pharmacological Genetics Research. Am. J. Hum. Genet. 83, 347–

358 (2008).

27. Patterson, N. et al. Ancient Admixture in Human History. Genetics 192, 1065–1093

(2012).

28. Ferrer-Admetlla, A. et al. A Natural History of FUT2 Polymorphism in Humans. Mol. Biol.

Evol. 26, 1993–2003 (2009).


https://doi.org/10.1101/055855

18

29. Fumagalli, M. et al. Widespread balancing selection and pathogen-driven selection at

blood group antigen genes. Genome Res. 19, 199–212 (2009).

30. Thorven, M. et al. A Homozygous Nonsense Mutation (428G→A) in the Human Secretor

(FUT2) Gene Provides Resistance to Symptomatic Norovirus (GGII) Infections. J. Virol. 79,

15351–15355 (2005).

31. Carlsson, B. et al. The G428A Nonsense Mutation in FUT2 Provides Strong but Not

Absolute Protection against Symptomatic GII.4 Norovirus Infection. PLOS ONE 4, e5593

(2009).

32. Kindberg, E. et al. A nonsense mutation (428G→A) in the fucosyltransferase FUT2 gene

affects the progression of HIV-1 infection: AIDS 20, 685–689 (2006).

33. Hazra, A. et al. Common variants of FUT2 are associated with plasma vitamin B12 levels.

Nat. Genet. 40, 1160–1162 (2008).

34. McGovern, D. P. B. et al. Fucosyltransferase 2 (FUT2) non-secretor status is associated

with Crohn’s disease. Hum. Mol. Genet. 19, 3468–3476 (2010).

35. Parmar, A. S. et al. Association study of FUT2 (rs601338) with celiac disease and

inflammatory bowel disease in the Finnish population. Tissue Antigens 80, 488–493 (2012).

36. Tong, M. et al. Reprograming of gut microbiome energy metabolism by the FUT2

Crohn’s disease risk polymorphism. ISME J. 8, 2193–2206 (2014).

37. Banda, Y. et al. Characterizing Race/Ethnicity and Genetic Ancestry for 100,000 Subjects

in the Genetic Epidemiology Research on Adult Health and Aging (GERA) Cohort. Genetics

200, 1285–1295 (2015).


https://doi.org/10.1101/055855

19

38. Renné, T., Schmaier, A. H., Nickel, K. F., Blombäck, M. & Maas, C. In vivo roles of factor

XII. Blood 120, 4296–4303 (2012).

39. Tang, W. et al. Genetic Associations for Activated Partial Thromboplastin Time and

Prothrombin Time, their Gene Expression Profiles, and Risk of Coronary Artery Disease. Am.

J. Hum. Genet. 91, 152–162 (2012).

40. Innocenti, F. et al. Identification, Replication, and Functional Fine-Mapping of

Expression Quantitative Trait Loci in Primary Human Liver Tissue. PLoS Genet. 7, (2011).

41. Guerrero, J. A. et al. Novel loci involved in platelet function and platelet count identified

by a genome-wide study performed in children. Haematologica 96, 1335–1343 (2011).

42. Wooding, S. P. et al. DNA Sequence Variation in a 3.7-kb Noncoding Sequence 5′ of the

CYP1A2 Gene: Implications for Human Population History and Natural Selection. Am. J. Hum.

Genet. 71, 528–542 (2002).

43. Ding, K. & Kullo, I. J. Geographic differences in allele frequencies of susceptibility SNPs

for cardiovascular disease. BMC Med. Genet. 12, 55 (2011).

44. Newton-Cheh, C. et al. Genome-wide association study identifies eight loci associated

with blood pressure. Nat. Genet. 41, 666–676 (2009).

45. Tabara, Y. et al. Common Variants in the ATP2B1 Gene Are Associated With

Susceptibility to Hypertension The Japanese Millennium Genome Project. Hypertension 56,

973–980 (2010).

46. Martin, J.-E. et al. Identification of CSK as a systemic sclerosis genetic risk factor through

Genome Wide Association Study follow-up. Hum. Mol. Genet. 21, 2825–2835 (2012).


https://doi.org/10.1101/055855

20

47. Xie, Q. et al. Decision Forest Analysis of 61 Single Nucleotide Polymorphisms in a Case-

Control Study of Esophageal Cancer; a novel method. BMC Bioinformatics 6, 1–9 (2005).

48. Cornelis, M. C. et al. Genome-Wide Meta-Analysis Identifies Regions on 7p21 (AHR) and

15q24 (CYP1A2) As Determinants of Habitual Caffeine Consumption. PLoS Genet. 7, (2011).

49. Popat, R. A. et al. Coffee, ADORA2A, and CYP1A2: the caffeine connection in Parkinson’s

disease. Eur. J. Neurol. Off. J. Eur. Fed. Neurol. Soc. 18, 756–765 (2011).

50. Hong, K.-W. et al. Recapitulation of two genomewide association studies on blood

pressure and essential hypertension in the Korean population. J. Hum. Genet. 55, 336–341

(2010).

51. Genetic Variants in Novel Pathways Influence Blood Pressure and Cardiovascular

Disease Risk. Nature 478, 103–109 (2011).

52. Sabeti, P. C. et al. Positive Natural Selection in the Human Lineage. Science 312, 1614–

1620 (2006).

53. Patterson, N., Price, A. L. & Reich, D. Population Structure and Eigenanalysis. PLoS Genet

2, e190 (2006).

54. Novembre, J. et al. Genes mirror geography within Europe. Nature 456, 98–101 (2008).

55. Lawson, D. J., Hellenthal, G., Myers, S. & Falush, D. Inference of Population Structure

using Dense Haplotype Data. PLoS Genet 8, e1002453 (2012).

56. The UK10K Consortium. The UK10K project identifies rare variants in health and disease.

Nature 526, 82–90 (2015).

57. Loh, P., Palamara, P. F. & Price, A. L. Fast and accurate long-range phasing in a UK

Biobank cohort. Nat. Genet. in press. http://biorxiv.org/content/early/2015/10/04/028282


https://doi.org/10.1101/055855

21

58. O’Connell, J. R., Sharp, K., Delaneau, O. & Marchini, J. Haplotype estimation for biobank

scale datasets. Nat. Genet. accepted in principle.

59. Pritchard, J. K., Pickrell, J. K. & Coop, G. The Genetics of Human Adaptation: Hard

Sweeps, Soft Sweeps, and Polygenic Adaptation. Curr. Biol. 20, R208–R215 (2010).

60. Pritchard, J. K. & Di Rienzo, A. Adaptation – not by sweeps alone. Nat. Rev. Genet. 11,

665–667 (2010).

61. Turchin, M. C. et al. Evidence of widespread selection on standing variation in Europe at

height-associated SNPs. Nat. Genet. 44, 1015–1019 (2012).

62. Berg, J. J. & Coop, G. A Population Genetic Signal of Polygenic Adaptation. PLOS Genet

10, e1004412 (2014).

63. Robinson, M. R. et al. Population genetic differentiation of height and body mass index

across Europe. Nat. Genet. 47, 1357–1362 (2015).

64. Yang, W.-Y., Novembre, J., Eskin, E. & Halperin, E. A model-based approach for analysis

of spatial structure in genetic data. Nat. Genet. 44, 725–731 (2012).

65. Baran, Y., Quintela, I., Carracedo, Á., Pasaniuc, B. & Halperin, E. Enhanced Localization of

Genetic Samples through Linkage-Disequilibrium Correction. Am. J. Hum. Genet. 92, 882–894

(2013).

66. Baran, Y. & Halperin, E. A Note on the Relations Between Spatio-Genetic Models. J.

Comput. Biol. 22, 905–917 (2015).

67. Chang, C. C. et al. Second-generation PLINK: rising to the challenge of larger and richer

datasets. GigaScience 4, 7 (2015).


https://doi.org/10.1101/055855

22

68. Heffelfinger, C. et al. Haplotype structure and positive selection at TLR1. Eur. J. Hum.

Genet. 22, 551–557 (2014).

69. Burton, P. R. et al. Genome-wide association study of 14,000 cases of seven common

diseases and 3,000 shared controls. Nature 447, 661–678 (2007).

70. Pickrell, J. K. et al. Signals of recent positive selection in a worldwide sample of human

populations. Genome Res. 19, 826–837 (2009).

71. de Bakker, P. I. W. et al. A high-resolution HLA and SNP haplotype map for disease

association studies in the extended human MHC. Nat. Genet. 38, 1166–1172 (2006).

72. Voight, B. F., Kudaravalli, S., Wen, X. & Pritchard, J. K. A Map of Recent Positive Selection

in the Human Genome. PLoS Biol 4, e72 (2006).


https://doi.org/10.1101/055855

23

Figures

Results of PCA with k-means clustering

The top 5 PCs in UK Biobank data are displayed. Samples were clustered using these PCs into 6

clusters with k-means clustering (see Table 1). PC5 is plotted against PC2, because PC5

primarily separated the orange and red clusters, which were separated from the other clusters

by PC2.


https://doi.org/10.1101/055855

24

Results of PCA with projection of PoBI samples

The top 5 PCs in UK Biobank data are displayed with PoBI samples projected onto these PCs.

PoBI populations which visually best matched the clusters from k-means clustering were used

to assign names to the six clusters (Table 1).


https://doi.org/10.1101/055855

25

Results of PCA with projection of ancient samples

The top 5 PCs in UK Biobank data are displayed with ancient samples projected onto these PCs.


https://doi.org/10.1101/055855

26

Selection statistics for UK Biobank along PC1

A Manhattan plot with − log10(𝑝) values is displayed. Values above the significance threshold

(dotted line, 𝑝 = 1.96 × 10−8, 𝛼 = 0.05 after correcting for 5 PCs and 510,665 SNPs) are

displayed as larger points and are labeled with the locus they correspond to (see Table 3).

− log10(𝑝) values larger than 10 are truncated at 10 for easier visualization and are displayed

as even larger points.


https://doi.org/10.1101/055855

27

Tables

Table 1 Correspondence between UK Biobank clusters and PoBI populations

We report the PoBI population that most closely corresponds to each UK Biobank cluster (see

main text).

Color Count Cluster Name PoBI Populations

Blue 41,494 Southern England Hampshire, Devon, Norfolk

Purple 19,452 Northern England Yorkshire, Lancashire

Brown 12,895 Northern Ireland Northern Ireland

Green 21,215 Scotland Argyll and Bute, Banff and Buchan, Orkney

Red 14,190 North Wales North Wales

Orange 4,605 South Wales / Pembrokeshire North Pembrokeshire, South Pembrokeshire


https://doi.org/10.1101/055855

28

Table 2 Results of f4 statistics in ancient and modern British samples

We report f4 statistics of of the form 𝑓4(𝑆𝑡𝑒𝑝𝑝𝑒, 𝑁𝑒𝑜𝑙𝑖𝑡ℎ𝑖𝑐 𝐹𝑎𝑟𝑚𝑒𝑟; 𝑃𝑜𝑝1, 𝑃𝑜𝑝2),

representing a z-score with positive values indicating more Steppe ancestry in Pop1 than Pop2 .

Samples for Pop1 were either modern Celtic (Scotland and Wales) or ancient Saxon. Samples for

Pop2 were modern Anglo-Saxon (southern and eastern England).

Pop2

Grouping Pop1 Hampshire Devon Norfolk

Ancient Saxon 2.543 3.732 5.118

Scotland Argyll and Bute 3.323 6.223 9.560

North Wales North Wales 1.918 5.239 8.490

South Wales North Pembrokeshire 1.759 4.430 7.124


https://doi.org/10.1101/055855

29

Table 3 Top signals of selection for UK Biobank along PC1-PC5

We report the top signal of natural selection for each locus reaching genome-wide significance

(𝑝 < 1.96 × 10−8) along any of the top five PCs. Neighboring SNPs <1Mb apart with genome-

wide significant signals were grouped together into a single locus.

Locus Chromosome Position (Mb) PC Top SNP p-value

LCT6 2 134.9 - 137.2 1 rs7570971 3.96 × 10−15 TLR168 4 38.8 - 38.9 1 rs4833095 7.96 × 10−15 2 1.27 × 10−8 3 7.89 × 10−9 4 1.54 × 10−11 IRF469,70 6 0.4 - 0.5 1 rs62389423 2.31 × 10−43

HLA71 6 31.1 - 32.9 1 rs9366778 8.45 × 10−9 FUT2 19 49.2 - 49.2 1 rs601338 9.16 × 10−9


https://doi.org/10.1101/055855

30

Table 4 Top signals of selection for combined selection statistics

We report the top selection statistic for each locus reaching genome-wide significance,

restricting to loci that were not genome-wide significant in either the UK Biobank selection

statistic or the ancient Eurasian selection statistics. Neighboring SNPs <1Mb apart with

genome-wide significant signals were grouped together into a single locus.

Locus Chr Position (Mb) PC Top SNP

Combined p-value

UK Biobank p-value

Ancient Eurasian p-value

F12 5 33.9 - 34.0 2 rs2545801 1.79 × 10−9 4.36 × 10−4 5.35 × 10−8 CYP1A2 / CSK

15 75.0 - 75.1 2 rs1378942 4.65 × 10−8 1.05 × 10−2 1.08 × 10−7


https://doi.org/10.1101/055855

31

Supplementary Figures

Supplementary Figure 1 Results of PCA after removing long-range LD regions

Regions with high SNP weights from the first PCA run were removed and PCA was run on the

remainder of the genome (see Online Methods). The resulting PCs are no longer influenced by

long-range LD regions. A visual inspecion suggests that PC1-PC5 have interesting population

structure while PC6-PC10 do not.


https://doi.org/10.1101/055855

32

Supplementary Figure 2 Results of PCA with k-means clustering for all PCs

This is an expanded set of plots similar to Figure 1, except that plots of all pairs of top PCs are

displayed.


https://doi.org/10.1101/055855

33

Supplementary Figure 3 Results of PCA with projection of PoBI samples for all PCs


displayed.


https://doi.org/10.1101/055855

34

Supplementary Figure 4 Results of PCA with projection of ancient samples for all PCs


displayed.


https://doi.org/10.1101/055855

35

Supplementary Figure 5 Results of PCA with projection of POPRES samples for all PCs

This set of plots is similar to Supplementary Figure 2, except that POPRES samples are projected

on top.


https://doi.org/10.1101/055855

36

Supplementary Figure 6 Selection statistic for UK Biobank along PC1-PC5

This is an expanded set of plots similar to Figure 4, except that plots for each of the top 5 PCs

are displayed.


https://doi.org/10.1101/055855

37

Supplementary Figure 7 Results of initial PCA run

This set of plots is similar to Supplementary Figure 1, except that long-range LD regions were

not removed. Several of these PCs are dominated by regions of long-range LD. In particular,

the three clusters along PC2 indicate 0, 1 or 2 copies of a chromosome 8 inversion variant.


https://doi.org/10.1101/055855

38

Supplementary Tables

Supplementary Table 1 PC eigenvalues and geographical correlations

PC1-PC5 all had elevated eigenvalues, while PC6-PC10 had eigenvalues which were close to

background levels. In the case where there are two equal-sized sample sets from distinct

populations, the 𝐹𝑆𝑇 between the two populations can be estimated from the top eigenvalue

(𝜆) via the following formula: 𝐹𝑆𝑇 = (𝜆 − 1)/𝑁, where 𝑁 is the total number of samples. The

top eigenvalue reflects an 𝐹𝑆𝑇 of 1.76 × 10−4, indicating very subtle population structure

within the UK. PC1 was most strongly correlated with east-west birth coordinate and PC2 was

most strongly correlated with north-south birth coordinate.

East-West North-South

Eigenvalue 𝑭𝑺𝑻 Correlation p-value Correlation p-value

PC1 20.99 1.76E-04 0.4154 <1e-50 -0.3981 <1e-50 PC2 9.35 7.33E-05 -0.0865 <1e-50 -0.4322 <1e-50 PC3 7.76 5.94E-05 0.1894 <1e-50 -0.1262 <1e-50 PC4 5.18 3.68E-05 -0.1418 <1e-50 0.3409 <1e-50 PC5 5.13 3.63E-05 0.0019 5.27E-01 -0.0124 4.14E-05 PC6 4.62 3.18E-05 -0.0163 6.84E-08 0.0150 6.93E-07 PC7 4.61 3.17E-05 -0.0025 4.01E-01 0.0049 1.04E-01 PC8 4.59 3.15E-05 0.0216 7.75E-13 0.0047 1.16E-01

PC9 4.59 3.15E-05 -0.0522 <1e-50 -0.0119 7.92E-05 PC10 4.57 3.14E-05 -0.0143 2.23E-06 0.0121 6.47E-05


https://doi.org/10.1101/055855

39

Supplementary Table 2 Expanded results of f4 statistics in ancient and modern British

samples

We report 𝑓4 statistics of of the form 𝑓4(𝑆𝑡𝑒𝑝𝑝𝑒, 𝑁𝑒𝑜𝑙𝑖𝑡ℎ𝑖𝑐 𝐹𝑎𝑟𝑚𝑒𝑟; 𝑃𝑜𝑝1, 𝑃𝑜𝑝2),

representing a z-score with positive values indicating more Steppe ancestry in Pop1 than Pop2 .

Samples for Pop1 were either modern Celtic (Scotland and Wales) or ancient Saxon. Samples for

Pop2 were modern Anglo-Saxon (southern and eastern England).

Pop2

Grouping Pop1 Norfolk Suffolk Hampshire Kent Devon

Saxon Saxon 5.118 5.268 2.543 3.953 3.32

Scotland Argyll and Bute 9.560 9.370 3.323 6.411 6.223

Banff and Buchan 7.609 77.545 1.234 4.440 4.379

Orkney 11.229 10.583 3.620 7.310 7.259

N. Wales North Wales 8.490 8.393 1.918 5.163 5.239

S. Wales North Pembrokeshire 7.124 7.287 1.759 4.542 4.430

South Pembrokeshire 6.301 6.189 2.315 4.336 4.171


https://doi.org/10.1101/055855

40

Supplementary Table 3 Suggestive signals of selection in UK Biobank

We report the top signal of natural selection for each locus not reaching genome-wide

significance (𝑝 > 1.96 × 10−8) but yielding a suggestive signal (𝑝 < 1.00 × 10−6) along any of

the top five PCs. Neighboring SNPs <1Mb apart with suggestives significant signals were

grouped together into a single locus.

Annotation Chromosome Locus (Mb) PC Best hit p-value

1 208.8 - 208.8 2 rs75602597 9.71e-07

ABCD3 1 226.4 - 226.8 4 rs72759068 8.62e-08

4 45.2 - 45.2 1 rs77147311 9.78e-07

5 164.8 - 164.9 2 rs77635680 2.13e-08

ZDHHC14 6 158.1 - 158.1 4 rs73584091 5.46e-07

7 64.9 - 64.9 2 rs79415723 8.81e-07

7 118.2 - 118.2 1 rs187417794 3.66e-07

13 66.2 - 66.2 3 rs1417218 8.38e-07

OCA270,72 15 28.4 - 28.4 2 rs12913832 5.55e-08

RPGRIP1L 16 53.7 - 53.7 2 rs61747071 7.81e-07

BPIFB9P 20 31.8 - 32.0 4 rs293709 3.00e-07

20 39.1 - 39.1 1 rs2143877 5.03e-07


https://doi.org/10.1101/055855

41

Supplementary Table 4 Allele frequency of FUT2 alleles

We report the allele frequency of the most significant hit, rs601338, along with two other

linked SNPs in GERA and the PoBI datasets.

Dataset Cluster rs601338 (G/A)

rs492602 (G/A)

rs676388 (T/C)

UK Biobank Northern Ireland 0.4406 0.4413 0.4207

Northeast England 0.4633 0.4638 0.4407

Pembrokeshire 0.4864 0.4871 0.4580

North Wales 0.5006 0.501 0.4781

Yorkshire 0.5025 0.503 0.4763

East Anglia 0.5109 0.5111 0.4847

GERA Irish 0.4754 0.4522

Northern European 0.5215 0.4962

Southern European 0.5248 0.5021

Ashkenazi Jewish 0.5530 0.5200

Eastern European 0.5840 0.5586

PoBI Argyll and Bute 0.2949

North Pembrokeshire 0.3171

Banff and Buchan 0.3365

Northern Ireland 0.3667

Cumbria 0.4039

Derbyshire 0.4091

Dorset 0.4125

Herefordshire 0.4259

Worcestershire 0.4265

Lancashire 0.4306

Devon 0.4416

Yorkshire 0.4517

Orkney 0.4531

Lincolnshire 0.4619

Kent 0.4661

Suffolk 0.4699

Cornwall 0.4716

Leicestershire 0.4726

North Wales 0.4737

Northeast England 0.4844

Cheshire 0.4875

Forest of Dean 0.4881

Norfolk 0.4951


https://doi.org/10.1101/055855

42

Nottinghamshire 0.5000

Northamptonshire 0.5000

Oxfordshire 0.5054

Sussex 0.5167

Gloucestershire 0.5417

South Pembrokeshire 0.5417

Hampshire 0.5833


https://doi.org/10.1101/055855

43

Supplementary Table 5 Discrete test for natural selection at FUT2 in GERA

We report results of tests for selection using discrete subpopulations for FUT2 in the GERA data

set. FUT2 does not reach genome-wide significance in the GERA dataset, however there are

several suggestive signals when comparing the “Irish” subgroup with the Northern European

and Eastern European subpopulation.

𝑭𝑺𝑻 EE IR NE SE

AJ 6.84E-03 6.71E-03 6.54E-03 3.45E-03

EE 9.44E-04 7.23E-04 2.39E-03

IR 1.26E-04 1.91E-03

NE 1.80E-03

rs492602 EE IR NE SE

AJ 5.96E-01 1.84E-01 5.83E-01 5.05E-01

EE 1.48E-06 1.58E-03 9.25E-02

IR 1.34E-07 1.16E-01

NE 9.12E-01

rs676388 EE IR NE SE

AJ 5.13E-01 2.45E-01 6.79E-01 6.73E-01

EE 2.49E-06 1.69E-03 1.10E-01

IR 4.53E-07 1.12E-01

NE 8.46E-01


https://doi.org/10.1101/055855

44

Supplementary Table 6 Discrete test for natural selection at FUT2 in UK Biobank

We report results of tests for selection using discrete subpopulations for FUT2 in the UK

Biobank data set, using the UK Biobank subpopulations derived from k-means clustering. With

15 comparisons per SNP and 510,665 SNPs (p-value threshold of 6.53 × 10−9), we are still able

to find genome-wide-significant results when comparing East Anglia with Northeast England as

well as Northern Ireland.

𝑭𝑺𝑻 2 3 4 5 6

1 7.46E-05 1.23E-04 8.53E-05 2.60E-04 2.23E-04

2 1.10E-04 7.36E-05 2.90E-04 2.67E-04

3 1.48E-04 3.10E-04 1.22E-04

4 3.75E-04 2.96E-04

5 3.43E-04

rs601338 2 3 4 5 6

1 1.42E-01 1.95E-05 7.94E-01 2.72E-01 1.42E-07

2 2.61E-09 2.22E-01 6.45E-02 6.04E-09

3 2.38E-05 9.12E-02 9.31E-03

4 2.77E-01 1.28E-06

5 1.47E-03

rs492602 2 3 4 5 6

1 1.51E-01 1.65E-05 8.49E-01 2.72E-01 1.65E-07

2 3.11E-09 2.29E-01 6.86E-02 7.35E-09

3 2.50E-05 8.93E-02 9.86E-03

4 2.85E-01 1.42E-06

5 1.48E-03

rs676388 2 3 4 5 6

1 2.91E-01 1.30E-05 9.82E-01 1.97E-01 3.28E-07

2 3.38E-08 2.13E-01 4.27E-02 1.07E-07

3 1.25E-04 2.06E-01 2.12E-02

4 2.17E-01 1.31E-05

5 9.38E-03


https://doi.org/10.1101/055855

45

Supplementary Table 7 Independence of UK Biobank and ancient Eurasian scans for

selection

The UK Biobank and ancient Eurasian selection statistics were not inflated genome-wide, nor at

the overlapping SNPs in both datasets. Similarly, the combined selection statistic was not

inflated either. The correlation between the two statistics is also small, with the UK Biobank

PC1 and ancient Eurasian statistics being most correlated with 𝑟 = 0.188.

Inflation

Genome-wide Overlap Combined Correlation

Ancient 1.00 1.07

UKB PC1 1.02 1.08 1.06 18.8%

UKB PC2 0.95 1.00 1.05 2.8%

UKB PC3 0.95 1.00 1.05 6.5%

UKB PC4 0.88 0.94 1.04 2.5%

UKB PC5 0.86 0.91 1.04 0.0%


https://doi.org/10.1101/055855

46

Supplementary Table 8 Phenotype associations at SNPs with signals of selection

We tested SNPs with genome-wide significant signals of selection in the constituent UK Biobank

or ancient Eurasian scans or the combined scan for association with 15 phenotypes in the UK

Biobank data set, using the top 5 PCs as covariates.

Locus Chr Pos (Mb) Phenotype Top SNP p-value

LCT 2 134.9 - 137.0 lung_FVCzSMOKE rs6716536 4.90e-10

TLR1 4 38.8 - 38.9 disease_ALLERGY_ECZEMA_DIAGNOSED rs5743614 5.80e-26

SLC22A4 5 131.4 - 131.8 body_HEIGHTz rs1050152 3.70e-18

disease_ASTHMA_DIAGNOSED rs2188962 6.00e-09

F12 5 176.8 - 176.8 body_HEIGHTz rs2545801 4.80e-11

HLA 6 28.3 - 33.0 body_HEIGHTz rs2256183 4.10e-23

body_WHRadjBMIz rs521977 1.80e-10

bp_DIASTOLICadjMEDz rs521977 2.00e-10

disease_ALLERGY_ECZEMA_DIAGNOSED rs3135377 1.60e-11

disease_ASTHMA_DIAGNOSED rs204993 5.10e-09

lung_FEV1FVCzSMOKE rs3891175 3.50e-21

lung_FVCzSMOKE rs6456834 3.70e-09

FADS1 11 61.5 - 61.6 bmd_HEEL_TSCOREz rs174548 1.20e-08

ATXN2/SH2B3 12 111.9 - 112.9 bp_DIASTOLICadjMEDz rs3184504 8.00e-33

bp_SYSTOLICadjMEDz rs3184504 1.90e-13

disease_HYPERTENSION_DIAGNOSED rs3184504 1.30e-09

CYP1A2/CSK 15 75.0 - 75.1 bp_DIASTOLICadjMEDz rs2472304 1.10e-19

bp_SYSTOLICadjMEDz rs2472304 4.20e-10

disease_HYPERTENSION_DIAGNOSED rs2472304 2.60e-09


https://doi.org/10.1101/055855

47

Supplementary Table 9 PC-phenotype associations in UK Biobank

We report the results of tests of associations (p-value) between top PCs and 15 phenotypes in

UK Biobank. This analysis does not distinguish between environmental and genetic effects.

Phenotype PC1 PC2 PC3 PC4 PC5

bmd_HEEL_TSCOREz 2.33E-01 4.75E-01 2.39E-23 3.54E-07 2.97E-01

body_BMIz 2.36E-27 7.98E-01 2.59E-11 2.21E-03 1.14E-01

body_HEIGHTz <1e-50 4.66E-01 <1e-50 8.38E-03 2.41E-03

body_WHRadjBMIz 2.75E-06 3.35E-01 3.71E-03 2.94E-24 1.06E-01

bp_DIASTOLICadjMEDz 7.34E-01 5.42E-01 2.70E-08 6.16E-13 2.37E-01

bp_SYSTOLICadjMEDz 1.75E-03 7.67E-02 6.31E-01 6.15E-07 3.09E-02

cov_EDU_COLLEGE 6.73E-01 9.75E-25 2.01E-29 8.59E-36 7.96E-07

cov_SMOKING_STATUS 7.13E-01 5.67E-01 3.61E-03 9.63E-07 8.08E-01

disease_ALLERGY_ECZEMA_DIAGNOSED 1.76E-16 1.50E-03 1.07E-08 7.15E-03 7.87E-01

disease_ASTHMA_DIAGNOSED 7.04E-01 2.50E-06 6.12E-01 2.89E-05 3.84E-01

disease_HYPERTENSION_DIAGNOSED 7.84E-01 2.45E-01 1.09E-03 5.92E-01 3.32E-01

lung_FEV1FVCzSMOKE 9.85E-01 3.31E-07 1.89E-13 5.58E-10 3.43E-07

lung_FVCzSMOKE 8.69E-02 1.93E-45 2.21E-03 3.48E-40 3.99E-04

repro_MENARCHE_AGE 1.14E-04 1.11E-05 1.93E-03 3.11E-01 3.75E-02

repro_MENOPAUSE_AGE 1.46E-15 7.51E-04 9.54E-07 1.27E-03 1.93E-01


https://doi.org/10.1101/055855

48

Supplementary Table 10 Significant or suggestive signals of selection in initial PCA run

We report significant or suggestive signals of selection in the initial PCA run. Neighboring SNPs

<1Mb apart with genome-wide significant signals were grouped together into a single locus.

The significant signals may represent either signals of selection or regions of long-range LD. All

of these regions were removed from the main PCA run (see Online Methods).

Chrom Locus (Mb) PC Best hit p-value

1 2.2 - 2.2 3 rs79907870 4.68E-07

1 54.8 - 54.8 1 rs17390412 9.13E-07 1 56.0 - 56.1 8 rs1875068 1.86E-08 2 88.7 - 88.7 4 rs1713939 1.02E-07 2 133.3 - 144.0 1 rs7570971 7.21E-18 4 rs1446585 3.02E-14 7 rs1446585 3.09E-13

10 rs72847650 <1e-50

2 159.8 - 160.0 7 rs1522699 3.89E-07 2 223.9 - 223.9 7 rs1900725 2.71E-07 3 46.3 - 46.4 7 rs9990343 2.80E-08 4 23.3 - 23.3 3 rs114557362 2.95E-07 4 38.7 - 38.9 1 rs4833095 9.29E-16

3 rs4833095 8.54E-11 4 rs4833095 1.21E-10 7 rs4833095 2.18E-16 5 60.6 - 60.6 3 rs10471511 6.04E-07

5 101.5 - 101.6 8 rs411954 1.20E-08 5 114.8 - 114.8 8 rs895291 5.87E-07

5 164.8 - 164.9 3 rs77635680 6.70E-10 6 0.4 - 0.7 1 rs62389423 1.29E-47 7 rs62389423 1.57E-09 6 23.9 - 36.7 1 rs151341075 3.76E-11 3 rs2253908 5.43E-07 4 rs151341075 3.35E-17

5 rs3131618 <1e-50 6 rs204999 <1e-50

7 rs2596573 3.27E-54 8 rs41268932 2.58E-23 9 rs2596573 <1e-50 10 rs9266258 4.14E-07


https://doi.org/10.1101/055855

49

6 46.8 - 46.8 7 rs9395218 9.92E-07 6 86.0 - 87.0 10 rs2816583 2.23E-08

7 41.4 - 41.4 1 rs76920365 3.66E-07 7 64.4 - 66.4 3 rs79415723 1.17E-08 7 118.2 - 118.3 1 rs187417794 1.13E-07 8 7.2 - 12.7 2 rs11250099 <1e-50 9 14.0 - 14.0 3 rs12380860 4.45E-07 10 133.2 - 133.2 4 rs57105422 7.45E-07

15 28.4 - 28.4 3 rs12913832 9.17E-10

15 50.8 - 50.8 5 rs148783236 3.18E-194 6 rs148783236 <1e-50 8 rs148783236 2.18E-13 9 rs148783236 <1e-50

16 9.5 - 9.5 4 rs12149526 6.26E-07

16 26.5 - 26.5 3 rs73528772 4.01E-07 16 53.7 - 53.7 3 rs61747071 1.36E-07 16 89.7 - 89.8 4 rs449882 5.37E-07 17 29.6 - 29.6 3 rs11655238 5.98E-07 19 33.8 - 33.8 3 rs41355649 3.65E-08

19 49.2 - 50.2 1 rs601338 1.05E-09

20 39.1 - 39.1 1 rs2143877 8.34E-08 22 32.9 - 32.9 5 rs115815765 7.40E-31 6 rs115815765 <1e-50


https://doi.org/10.1101/055855

population structure of uk biobank and ancient eurasians reveals ...€¦ · 27/05/2016 · we...

Documents